Hi all,
I live in India. Here ambient tempretures easily reach 40deg C+ often a bit higher. This is coupled with high humidity for about 6 months of the year (May-Oct).
I have been thinking of building a Aleph design but worry that with all the ambient heat they will not perform as they should and add even more heat to the listening room.
I like the concept of SE class A. The 2 SE amps (albeit low watt tube) amps I have listened to bith produced a very nice sound.
I wonder if Mr. Pass or anyone else out here has any design that is suitbale for enviroments where heat dissipation is a problem. Sort of like a Cool Class A SE (is this an oxymoron)? My need is for a 100W and a 50W amp. I intend to biamp my speaker with 100W for the bass and 50W for the mids and HF.
I live in India. Here ambient tempretures easily reach 40deg C+ often a bit higher. This is coupled with high humidity for about 6 months of the year (May-Oct).
I have been thinking of building a Aleph design but worry that with all the ambient heat they will not perform as they should and add even more heat to the listening room.
I like the concept of SE class A. The 2 SE amps (albeit low watt tube) amps I have listened to bith produced a very nice sound.
I wonder if Mr. Pass or anyone else out here has any design that is suitbale for enviroments where heat dissipation is a problem. Sort of like a Cool Class A SE (is this an oxymoron)? My need is for a 100W and a 50W amp. I intend to biamp my speaker with 100W for the bass and 50W for the mids and HF.
Chokes to the rescue!
My favourite approach to this problem - and fast becoming my trademark so it seems - is to use a source follower output stage with a 100mH or so choke in the source leg. Speaker is wired across the choke, maybe with a blocking cap depending on the choke dc voltage drop. This approach only needs half the supply rail voltage of an Aleph style class A cct and therefore has only half the dissipation and half sized power supply. My particular version has 27 and a bit volts supply rail and 3.5 amps quiescent per channel and makes almost 50w aside into 8 ohms. Sound is unexpectedly good. You may already have seen it but the thread is here. http://www.diyaudio.com/forums/showthread.php?s=&threadid=6338
My favourite approach to this problem - and fast becoming my trademark so it seems - is to use a source follower output stage with a 100mH or so choke in the source leg. Speaker is wired across the choke, maybe with a blocking cap depending on the choke dc voltage drop. This approach only needs half the supply rail voltage of an Aleph style class A cct and therefore has only half the dissipation and half sized power supply. My particular version has 27 and a bit volts supply rail and 3.5 amps quiescent per channel and makes almost 50w aside into 8 ohms. Sound is unexpectedly good. You may already have seen it but the thread is here. http://www.diyaudio.com/forums/showthread.php?s=&threadid=6338
As soon as you go Class A (real Class A, not sliding bias) and single ended, you have to kiss efficiency goodbye. That means most of your power is going into heating the room. It's a sad thing, but it's basic physics.
Is it 40 C in the house? We have the same temps here in Tucson, just not in my listening room 🙂
Russ
Russ
Navin:
Your case (high ambient air temps) might be one of the ones that really would benefit from the idea of water cooling. It would be a fun little engineering project, and I think not too difficult. Still, in the end, the heat must be transferred to the air since you would want to recycle the water in a closed-loop system. But with something like a small automotive radiator or oil cooler and a fan that runs off the mains, it would work well I think. A potential problem is where to put the thing. I would like to do this someday, and plan to put it in the crawlspace beneath my house.
Alternately, just outside a window, in a little micro-shed. Live in a flat? Hmmmm. Maybe in the kitchen under a counter?
Your case (high ambient air temps) might be one of the ones that really would benefit from the idea of water cooling. It would be a fun little engineering project, and I think not too difficult. Still, in the end, the heat must be transferred to the air since you would want to recycle the water in a closed-loop system. But with something like a small automotive radiator or oil cooler and a fan that runs off the mains, it would work well I think. A potential problem is where to put the thing. I would like to do this someday, and plan to put it in the crawlspace beneath my house.
Alternately, just outside a window, in a little micro-shed. Live in a flat? Hmmmm. Maybe in the kitchen under a counter?
Maybe you could hook it up to the house's hot/cold water supply somehow, so it heats the water tank? 😉 (You got modern plumbing, right? 😉 )
If you can even just put the output devices out a window, then you won't have to deal with the heat indoors. And face it, SS can take 100°C. Don't worry about a 40°C temp rise...
An added plus to having the heatsink outdoors is it'll stay real cool during monsoon season. 🙂 🙂
I agree with Circ's output choke idea, much more efficiency...(such that it still is 😉 ).
Tim
If you can even just put the output devices out a window, then you won't have to deal with the heat indoors. And face it, SS can take 100°C. Don't worry about a 40°C temp rise...
An added plus to having the heatsink outdoors is it'll stay real cool during monsoon season. 🙂 🙂
I agree with Circ's output choke idea, much more efficiency...(such that it still is 😉 ).
Tim
i'd love to use the great outdoors to cool the amps.
Q1: the transistors are normally attached to teh heatsinks. by water cooling these heatsinks (either using recirculation or rain) what happens if these transistors get wet. Also what do I do with the heat from the trnasformer or is this not significant. We also have 90%+ humidity for 6 months. This humidity can settle on the transistors. It settles on furniture.
Q2: I live in an apt. In bombay everyone but real big and old money does. A lower watt version of the Aleph would sufffice. Is there one?
Q3: My stereo as to double as AV/HT too. Given that I will be locating the power amps somewhere else can i build 10 channels in one box, fed from a single large transformer but having seperate rectifiers and caps? I hope my front 3 channels will be a biamped 3 way and the 4 rears will be a 2 way. 50W per amp is adequate.
I am looking a a big 3KVA transformer which will feed 10 seperate bridge rectifiers and 6x10,000uf of caps per channel.
Q1: the transistors are normally attached to teh heatsinks. by water cooling these heatsinks (either using recirculation or rain) what happens if these transistors get wet. Also what do I do with the heat from the trnasformer or is this not significant. We also have 90%+ humidity for 6 months. This humidity can settle on the transistors. It settles on furniture.
Q2: I live in an apt. In bombay everyone but real big and old money does. A lower watt version of the Aleph would sufffice. Is there one?
Q3: My stereo as to double as AV/HT too. Given that I will be locating the power amps somewhere else can i build 10 channels in one box, fed from a single large transformer but having seperate rectifiers and caps? I hope my front 3 channels will be a biamped 3 way and the 4 rears will be a 2 way. 50W per amp is adequate.
I am looking a a big 3KVA transformer which will feed 10 seperate bridge rectifiers and 6x10,000uf of caps per channel.
If your transistors are above the dewpoint, obviously condensation won't be a problem. Even at 100% humidity - by definition - dewpoint can't go above ambient (if it did, the air would turn to fog).
Water could be a problem, but that's why you do a quality job. 😉 Could have the amp stuck in a window, with the heatsink pointing out the back. A sloped top would keep rain outside and direct it onto the heatsink. Use plenty of silicone to seal the joints to keep the water out.
Even so, if it does get in, it shouldn't do much. Rain water has little conductivity, so it won't short out any but the highest impedance circuits. (I don't know anything about the Aleph but I can't imagine it using anything greater than 10kohms, in which case rain won't load it too much. If the amp is built to run a little hot, it will also dry out quickly. 😉 )
Hmm.. probably wouldn't be a bad idea to put some vents on the indoors side, so the moisture can get out, if some does get in. 🙂
Good luck!
Tim
Water could be a problem, but that's why you do a quality job. 😉 Could have the amp stuck in a window, with the heatsink pointing out the back. A sloped top would keep rain outside and direct it onto the heatsink. Use plenty of silicone to seal the joints to keep the water out.
Even so, if it does get in, it shouldn't do much. Rain water has little conductivity, so it won't short out any but the highest impedance circuits. (I don't know anything about the Aleph but I can't imagine it using anything greater than 10kohms, in which case rain won't load it too much. If the amp is built to run a little hot, it will also dry out quickly. 😉 )
Hmm.. probably wouldn't be a bad idea to put some vents on the indoors side, so the moisture can get out, if some does get in. 🙂
Good luck!
Tim
"Vacuum tube" ;-)
How about if you had the entire amp enclosed in a case, heatsinks and all, and had a single air exhaust hole with a length of 100mm diameter circular flexible ducting leading away to outside, with a fan sucking the air out at the far end so you can't hear it where the action is. Sort of like a large diameter vacuum cleaner.
How about if you had the entire amp enclosed in a case, heatsinks and all, and had a single air exhaust hole with a length of 100mm diameter circular flexible ducting leading away to outside, with a fan sucking the air out at the far end so you can't hear it where the action is. Sort of like a large diameter vacuum cleaner.
Hi Navin
I'm in Karachi and the weather is not too dissimilar to Bombay's. I'm using an Aleph 3 clone with a slightly higher than normal bias with forced air cooling without any problems. Four computer fans run at six volts blowing air through four heatsinks. The fans are quite silent and only audible up close.
Regards, Hussain
I'm in Karachi and the weather is not too dissimilar to Bombay's. I'm using an Aleph 3 clone with a slightly higher than normal bias with forced air cooling without any problems. Four computer fans run at six volts blowing air through four heatsinks. The fans are quite silent and only audible up close.
Regards, Hussain
Air cooling
Hi Navin,
You can use a box type heatsink used by many older designs. You can make a square box , all sides with 4mm aluminum sheet. Then make fins extending from one face to the opposite face. Maybe 4 or 5 mm apart. Use 0.5mm aluminum or better 1mm for this. Then attach a 6 inch cooling fan at one end. The fans cost about Rs600- (US$12/-). Depending on the number and width of the fins you should have enough cooling that will keep the amp HOT ( 70 deg C?) but Stable. Maybe lateral MOSFET power devices. I bought some from a streetside shop in Mumbai a few years ago. How much do they cost there today ?
At 70 degrees C no moisture will settle on your hot devices because the dew point under hot Mumbai conditions is about 30 degC. On the other hand air cooling may be better because of the denser air ! Maybe.
It may not be necessary to have a class A amp for the sub. The sub will also benefit from a normal class AB amp due to its very low output impedance. This may not be too critical for the upper frequencies. You already know all of this.
Class A is great but I would compromise with sliding bias and its better efficiency. I want at least a 100 watts on each channel and living in a hot place ( 30~33 C in house) I don't want to add to the heat. I would target 150 watts as a minimum for the sub.
You may also want to look at active crossovers with 30 watt Class A amps ( x4) . They sound great , we have tried them out. You would never believe that the amps are only 30 watts.
Hot solutions for a hot country!
Cheers.
I took so long to write this piece that Circlotron has already posted the basic design concept . Check the web as I saw some pictures somewhere.
Hi Navin,
You can use a box type heatsink used by many older designs. You can make a square box , all sides with 4mm aluminum sheet. Then make fins extending from one face to the opposite face. Maybe 4 or 5 mm apart. Use 0.5mm aluminum or better 1mm for this. Then attach a 6 inch cooling fan at one end. The fans cost about Rs600- (US$12/-). Depending on the number and width of the fins you should have enough cooling that will keep the amp HOT ( 70 deg C?) but Stable. Maybe lateral MOSFET power devices. I bought some from a streetside shop in Mumbai a few years ago. How much do they cost there today ?
At 70 degrees C no moisture will settle on your hot devices because the dew point under hot Mumbai conditions is about 30 degC. On the other hand air cooling may be better because of the denser air ! Maybe.
It may not be necessary to have a class A amp for the sub. The sub will also benefit from a normal class AB amp due to its very low output impedance. This may not be too critical for the upper frequencies. You already know all of this.
Class A is great but I would compromise with sliding bias and its better efficiency. I want at least a 100 watts on each channel and living in a hot place ( 30~33 C in house) I don't want to add to the heat. I would target 150 watts as a minimum for the sub.
You may also want to look at active crossovers with 30 watt Class A amps ( x4) . They sound great , we have tried them out. You would never believe that the amps are only 30 watts.
Hot solutions for a hot country!
Cheers.
I took so long to write this piece that Circlotron has already posted the basic design concept . Check the web as I saw some pictures somewhere.
i prefer to use readymade heatsinks. what i get here are extruded aluminum heatsinks that are 8" wide the fins are about 2.5" high. These are available in 1 meter lenghts and I can cut them to any size and round the corners. I used 2 10" portions of this heatsink for my 200W+ Class AB amp based on a B&K ST140 circuit (only my amp uses 3 pairs of power transistors per channel and a slightly high bias). For further cooling 4 computer fans (AC 220V) turn on via a thermistor set to 55 deg C.
I noticed that the lower part of stereo cabinets are too low to store CDs etc so it is best to use this space to store power amps. I can have 4 stereo amps (8 channels) in a 2 x 2 array. 8 channels means I can go 7.1 if I dont biamp ot triamp or I can go 2.1 if I triamp.
I noticed that the lower part of stereo cabinets are too low to store CDs etc so it is best to use this space to store power amps. I can have 4 stereo amps (8 channels) in a 2 x 2 array. 8 channels means I can go 7.1 if I dont biamp ot triamp or I can go 2.1 if I triamp.
The neighbours.
Hi Navin,
You sure have lots of power at hand. Does it heat up your neighbours?
How does the B&K clone sound ?
Have you heard the Burmester Test CD?
Too many questions.
Cheers.
Hi Navin,
You sure have lots of power at hand. Does it heat up your neighbours?
How does the B&K clone sound ?
Have you heard the Burmester Test CD?
Too many questions.
Cheers.
well my neighbours are tolerant.
i have only had the chance to compare the amp to a few mid fi amps here. Adcom 555 was one a friend has one to drive his Klipsch Chorus, Carver M1.0t is another (I use this for my bass), another friend had a B&K ST140 and this was my bechmark; the load was a pair of Maggies (the smalll ones that sold for about $500). The home made amp sounded better than the Adcom and Carver.
The Adcom owner was surprised. WE trsted both aps against the Klipsch as my speaker were not ready at that time.
The B&K owner wanted me to build a clone for him but thenhe moved and we lost touch.
The Carver was also compared. The Load (for both amps) in this case was a bassreflex MTM using 2 8" Focal woofers and a Morel MDT33. The Carver had softer midbass and a bit less pressence. Hence I used the Carver for the bass.
i have only had the chance to compare the amp to a few mid fi amps here. Adcom 555 was one a friend has one to drive his Klipsch Chorus, Carver M1.0t is another (I use this for my bass), another friend had a B&K ST140 and this was my bechmark; the load was a pair of Maggies (the smalll ones that sold for about $500). The home made amp sounded better than the Adcom and Carver.
The Adcom owner was surprised. WE trsted both aps against the Klipsch as my speaker were not ready at that time.
The B&K owner wanted me to build a clone for him but thenhe moved and we lost touch.
The Carver was also compared. The Load (for both amps) in this case was a bassreflex MTM using 2 8" Focal woofers and a Morel MDT33. The Carver had softer midbass and a bit less pressence. Hence I used the Carver for the bass.
Navin-
In case you have any more interest, I'll tell you more about my thoughts about water cooling.
Those Aleph amps are relatively efficient for class A amps. They are about 30-35% efficient.
So if you want 10 channels at 50W each, you will dissipate about 1500 watts all the time.
That's a lot of heat. You don't want that in your apartment, and that is a lot of 'conventional' heatsinking needed. Particularly with your high ambient air temperatures. And those ambient temps also seem to call for some fan cooling which would be loud in the listening area.
So, the idea I have in mind is to use some tubing to carry the heat elsewhere (via water) from the amps. The amps themselves could be much smaller than if they were to have the heatsinking incorporated integrally at the amp.
In the amps, the semiconductor devices could be mounted on some heat exchangers that consist of simply a bar of copper, say 5mm thick by 20 wide, long enough for the semiconductors needed for one channel, with a piece of copper tubing brazed to the back side of the copper bar. Semiconductors screwed into tapped holes in the bar. They don't come into touch with the water.
Hanging outside your window you have a unit that consists of a sump (maybe a gallon tank (plastic jug or some such)), a small pump to move the water, and some sort of a water-to-air heat exhanger. I had in mind a small auto-type radiator or oil cooler because it probably could be picked up as junk much cheaper than the standard electronic heatsinks, yet they also are specifically water-to-air exchangers. Lastly you need a fan to blow air through the thing. That fan could be something used for industrial uses. Outdoors and running on 220V, you could really move some air without the concern about noise that you would have in the apartment. In the grand scheme of things it would still be a small fan. Small fans are available here surplus for less than $20, and I presume that they might be in India as well.
So the outdoor heat exchanger pumps the water from itself through some plastic tubing to the amps where it flows through the copper tubing extracting heat from the amp. Then it flows back to the heat exchanger where it's cooled before going back to the amps. Not much control is needed; maybe a thermal switch to turn the fan on and off to control the temperature.
That's it! If you have further interest, I can always draw a sketch.
Regarding the Aleph amps, the Aleph 30 is 30W/chan and the Aleph 5 is 60W/chan. The topology is the same, with only minor differences. It would be easy enough to tweak either to make it 50W/chan if you wanted.
Best of luck!
In case you have any more interest, I'll tell you more about my thoughts about water cooling.
Those Aleph amps are relatively efficient for class A amps. They are about 30-35% efficient.
So if you want 10 channels at 50W each, you will dissipate about 1500 watts all the time.
That's a lot of heat. You don't want that in your apartment, and that is a lot of 'conventional' heatsinking needed. Particularly with your high ambient air temperatures. And those ambient temps also seem to call for some fan cooling which would be loud in the listening area.
So, the idea I have in mind is to use some tubing to carry the heat elsewhere (via water) from the amps. The amps themselves could be much smaller than if they were to have the heatsinking incorporated integrally at the amp.
In the amps, the semiconductor devices could be mounted on some heat exchangers that consist of simply a bar of copper, say 5mm thick by 20 wide, long enough for the semiconductors needed for one channel, with a piece of copper tubing brazed to the back side of the copper bar. Semiconductors screwed into tapped holes in the bar. They don't come into touch with the water.
Hanging outside your window you have a unit that consists of a sump (maybe a gallon tank (plastic jug or some such)), a small pump to move the water, and some sort of a water-to-air heat exhanger. I had in mind a small auto-type radiator or oil cooler because it probably could be picked up as junk much cheaper than the standard electronic heatsinks, yet they also are specifically water-to-air exchangers. Lastly you need a fan to blow air through the thing. That fan could be something used for industrial uses. Outdoors and running on 220V, you could really move some air without the concern about noise that you would have in the apartment. In the grand scheme of things it would still be a small fan. Small fans are available here surplus for less than $20, and I presume that they might be in India as well.
So the outdoor heat exchanger pumps the water from itself through some plastic tubing to the amps where it flows through the copper tubing extracting heat from the amp. Then it flows back to the heat exchanger where it's cooled before going back to the amps. Not much control is needed; maybe a thermal switch to turn the fan on and off to control the temperature.
That's it! If you have further interest, I can always draw a sketch.
Regarding the Aleph amps, the Aleph 30 is 30W/chan and the Aleph 5 is 60W/chan. The topology is the same, with only minor differences. It would be easy enough to tweak either to make it 50W/chan if you wanted.
Best of luck!
thanks. i am getting a nice picture in my head. mount each transistor on a copper bar.
each copper bar is about 3mm apart and parallel to the others. so we have astring of copper bars. maybe 2 rows facing each other. assuming 17" for the amp we can get 16 transistors in a 8 x 2 array.
these copper bars are then brazed to a copper tube that carries water. sort of like the tubes in a AC condersor coil. Only I dont want a motor (noisy and expensive) to be used for recycling the water maybe convection currrents can be used. any ideas.
assuming i use Aelph 5s. I get 8 60W channels of power transistors all sharing one heatsink.
each copper bar is about 3mm apart and parallel to the others. so we have astring of copper bars. maybe 2 rows facing each other. assuming 17" for the amp we can get 16 transistors in a 8 x 2 array.
these copper bars are then brazed to a copper tube that carries water. sort of like the tubes in a AC condersor coil. Only I dont want a motor (noisy and expensive) to be used for recycling the water maybe convection currrents can be used. any ideas.
assuming i use Aelph 5s. I get 8 60W channels of power transistors all sharing one heatsink.
You've got my idea.
But a couple of things-
I would suggest at least one copper bar per channel. Operate them in parallel, and try to engineer the setup so that each parallel path presents a similar resistance to the incoming water flow to get good distribution. They should probably each have a small valve at their outlets so that the flow can be adjusted and to provide some resistance so that the pressure builds up at the amp to keep the tubes full. (I'm still assuming a pump here).
On that pump issue, maybe you could get get convection to work, I have to say, I don't have any experience with that. I had in mind some cheap pump. There are units made for small fountains and fish ponds that might apply. Here's one on ebay to look at: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=2312517165&category=519 It would be mounted together with the outdoor heat exchanger and its fan, so noise wouldn't be an issue.
None of this is my invention, as we used to use something like this in an industrial installation in my previous work. The only thing that is new is the diy implementation. I have the benefit that I've seen most of the likely bits here and there in my travels as I have been keeping something like this in mind for a while.
But a couple of things-
I would suggest at least one copper bar per channel. Operate them in parallel, and try to engineer the setup so that each parallel path presents a similar resistance to the incoming water flow to get good distribution. They should probably each have a small valve at their outlets so that the flow can be adjusted and to provide some resistance so that the pressure builds up at the amp to keep the tubes full. (I'm still assuming a pump here).
On that pump issue, maybe you could get get convection to work, I have to say, I don't have any experience with that. I had in mind some cheap pump. There are units made for small fountains and fish ponds that might apply. Here's one on ebay to look at: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=2312517165&category=519 It would be mounted together with the outdoor heat exchanger and its fan, so noise wouldn't be an issue.
None of this is my invention, as we used to use something like this in an industrial installation in my previous work. The only thing that is new is the diy implementation. I have the benefit that I've seen most of the likely bits here and there in my travels as I have been keeping something like this in mind for a while.
Efficiency
I was under the impression that transformer loaded circuits were 50% efficient and usual solid state amps in class A were 25% efficient.
How does the efficiency go up ?
Thanks.
Cheers
I was under the impression that transformer loaded circuits were 50% efficient and usual solid state amps in class A were 25% efficient.
How does the efficiency go up ?
Thanks.
Cheers
Basically the current source biasing the output stage is active, and gets a control signal from the output node of the amplifier that varies the current source.
Read the patent here:
http://patft.uspto.gov/netacgi/nph-...0&s1=5710522.WKU.&OS=PN/5710522&RS=PN/5710522
Read the patent here:
http://patft.uspto.gov/netacgi/nph-...0&s1=5710522.WKU.&OS=PN/5710522&RS=PN/5710522
- Status
- Not open for further replies.